{ "checked": "2026-09-25", "notice": "This is a catalogue manifest, not a download of upstream data or model weights. Resource terms and availability can change. Pin and review a specific release before training or redistribution.", "resources": [ { "id": "pubchem", "name": "PubChem", "type": "Dataset", "domain": "Molecules & assays", "use": "Learn chemical identifiers and molecular properties; assemble a documented subset for feature engineering.", "acquire": "Use PUG-REST for small queries or the official bulk-download service for larger selections.", "licence": "Source-specific terms; inspect contributor licences", "licenceUrl": "https://pubchem.ncbi.nlm.nih.gov/docs/data-sources", "sourceUrl": "https://pubchem.ncbi.nlm.nih.gov/", "accessUrl": "https://pubchem.ncbi.nlm.nih.gov/docs/downloads", "learning": "Retrieve a small set of CIDs, compare properties and build a provenance-aware CSV.", "snapshot": "Record CIDs, queried fields, retrieval time, responses and checksums.", "mirror": "Download a selected snapshot after checking source terms. PubChem itself is not a GitHub repository.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "chembl", "name": "ChEMBL", "type": "Dataset", "domain": "Bioactivity", "use": "Learn assay-aware data cleaning and molecular-property or activity prediction.", "acquire": "Download a named release or use the ChEMBL web services for a bounded query.", "licence": "CC BY-SA 3.0", "licenceUrl": "https://www.ebi.ac.uk/chembl/", "sourceUrl": "https://www.ebi.ac.uk/chembl/", "accessUrl": "https://www.ebi.ac.uk/training/online/courses/chembl-quick-tour/accessing-chembl-data/", "learning": "Compare measurements for one assay type; preserve units, qualifiers and assay context.", "snapshot": "Pin the release, query, selected identifiers and transformations.", "mirror": "A versioned data subset can be maintained with attribution and applicable share-alike obligations.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "opentargets", "name": "Open Targets Platform", "type": "Dataset", "domain": "Targets & disease", "use": "Learn evidence integration and transparent drug-target-disease prioritisation.", "acquire": "Use the documented API for exploration and release downloads for reproducible batch work.", "licence": "CC0 platform data; Apache 2.0 platform code", "licenceUrl": "https://platform-docs.opentargets.org/licence", "sourceUrl": "https://platform.opentargets.org/", "accessUrl": "https://platform-docs.opentargets.org/data-access", "learning": "Build a small disease-target evidence table and inspect which evidence supports each association.", "snapshot": "Pin the platform release and selected datasets, then record identifiers and file hashes.", "mirror": "Platform data can be snapshotted. Keep original-source provenance; check separate upstream assets individually.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "uniprot", "name": "UniProt", "type": "Dataset", "domain": "Proteins", "use": "Learn protein identifiers, sequence annotations and protein-feature preparation.", "acquire": "Use the website export or documented REST API to select a bounded protein set.", "licence": "CC BY 4.0 for copyrightable database content", "licenceUrl": "https://www.uniprot.org/help/copyright", "sourceUrl": "https://www.uniprot.org/", "accessUrl": "https://www.uniprot.org/api-documentation", "learning": "Compare reviewed sequence records and learn how identifiers map to external databases.", "snapshot": "Record release, accessions, sequence versions, retrieval date and file hashes.", "mirror": "Preserve attribution and version metadata when redistributing a selected dataset.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "pdb", "name": "RCSB PDB / wwPDB", "type": "Dataset", "domain": "Structures", "use": "Learn structural data formats, structure quality and protein-ligand representation.", "acquire": "Download selected archive entries or use RCSB search and data APIs.", "licence": "CC0 archive data; external integrated resources may differ", "licenceUrl": "https://www.rcsb.org/pages/policies", "sourceUrl": "https://www.rcsb.org/", "accessUrl": "https://www.rcsb.org/docs/programmatic-access/file-download-services", "learning": "Inspect an experimentally determined structure and its quality metadata before modelling.", "snapshot": "Record PDB identifiers, revision dates, file hashes and any external-resource dependencies.", "mirror": "Archive entries can be snapshotted; retain citations and check external-resource terms separately.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "clinicaltrials", "name": "ClinicalTrials.gov", "type": "Dataset", "domain": "Clinical evidence", "use": "Learn trial-registry data structures and build evidence maps from study metadata.", "acquire": "Use the official API or site exports. Registry records are not individual participant datasets.", "licence": "ClinicalTrials.gov terms and conditions apply", "licenceUrl": "https://clinicaltrials.gov/about-site/terms-conditions", "sourceUrl": "https://clinicaltrials.gov/", "accessUrl": "https://clinicaltrials.gov/data-api/api", "learning": "Map study status, eligibility and outcome definitions for one research question.", "snapshot": "Record NCT identifiers, last update dates, query and retrieval timestamp.", "mirror": "Review the stated terms before redistributing records; preserve source dates and avoid implied endorsement.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "chebi", "name": "ChEBI", "type": "Dataset", "domain": "Chemical ontology", "use": "Learn chemical entity classification and ontology-based data integration.", "acquire": "Download ontology or structure files from the official ChEBI downloads page.", "licence": "CC BY 4.0", "licenceUrl": "https://www.ebi.ac.uk/chebi/about", "sourceUrl": "https://www.ebi.ac.uk/chebi/", "accessUrl": "https://www.ebi.ac.uk/chebi/downloads", "learning": "Join chemical identifiers to ontology classes and inspect ambiguous mappings.", "snapshot": "Record release, file checksums, selected identifiers and mapping rules.", "mirror": "Keep attribution and the selected version with any redistributed subset.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null }, { "id": "rdkit", "name": "RDKit", "type": "Software", "domain": "Cheminformatics", "use": "Learn molecular parsing, fingerprints and descriptor generation. This is a toolkit, not a pretrained model.", "acquire": "Install a pinned release or clone the upstream repository and follow its build instructions.", "licence": "BSD 3-Clause", "licenceUrl": "https://github.com/rdkit/rdkit/blob/master/license.txt", "sourceUrl": "https://www.rdkit.org/docs/Overview.html", "accessUrl": "https://github.com/rdkit/rdkit", "learning": "Turn a small lawful SMILES dataset into descriptors and compare representations.", "snapshot": "Pin the release or commit, environment, input records and descriptor configuration.", "mirror": "Fork the software while retaining copyright and licence notices; data rights remain separate.", "forkUrl": "https://github.com/rdkit/rdkit/fork", "checked": "2026-09-25", "status": "Upstream resource", "localFork": "https://github.com/zodenode/openpharma-rdkit" }, { "id": "deepchem", "name": "DeepChem", "type": "Software", "domain": "Machine learning", "use": "Learn model training and benchmark workflows. Dataset and checkpoint terms remain separate from the software licence.", "acquire": "Use the documented tutorials with a pinned DeepChem version and environment.", "licence": "MIT software; check individual datasets and weights", "licenceUrl": "https://deepchem.readthedocs.io/en/latest/development_guide/licence.html", "sourceUrl": "https://github.com/deepchem/deepchem", "accessUrl": "https://github.com/deepchem/deepchem", "learning": "Run a simple molecular-property baseline and compare random and scaffold splits.", "snapshot": "Pin the commit, dataset release, split, seeds, dependency lock and model settings.", "mirror": "Forkable software. Retain notices and independently review each downloaded learning asset.", "forkUrl": "https://github.com/deepchem/deepchem/fork", "checked": "2026-09-25", "status": "Upstream resource", "localFork": "https://github.com/zodenode/openpharma-deepchem" }, { "id": "esm2", "name": "ESM-2 ยท 8M checkpoint", "type": "Model", "domain": "Protein representation", "use": "Learn protein embeddings and masked-residue prediction using an existing pretrained checkpoint.", "acquire": "Read the model card and obtain weights from the official model repository.", "licence": "MIT as stated on the model card", "licenceUrl": "https://huggingface.co/facebook/esm2_t6_8M_UR50D", "sourceUrl": "https://huggingface.co/facebook/esm2_t6_8M_UR50D", "accessUrl": "https://huggingface.co/facebook/esm2_t6_8M_UR50D/tree/main", "learning": "Compare embeddings for a small public protein set; document sequence limits and preprocessing.", "snapshot": "Pin the model revision and weights checksum alongside tokenizer and library versions.", "mirror": "Version the checkpoint manifest and retain its terms. Code forks do not automatically include model weights.", "forkUrl": null, "checked": "2026-09-25", "status": "Upstream resource", "localFork": null } ] }